Add Strands AgentCore Evaluation integration

- Add evaluation module with Strands integration
- Implement StrandsEvalsAgentCoreEvaluator for built-in and custom evaluators
- Add converter for Strands OTel spans to ADOT format
- Add CloudWatch span fetching for Runtime agents
- Include comprehensive unit tests and end-to-end tests
- Add real integration tests for evaluation API
- Add README with usage examples and API reference

Author: Aidan Daly

pyproject.toml

@@ -94,6 +94,7 @@ convention = "google"
 testpaths = [
     "tests"
 ]
+asyncio_mode = "auto"
 
 [tool.coverage.run]
 branch = true
@@ -148,9 +149,13 @@ dev = [
     "websockets>=14.1",
     "wheel>=0.45.1",
     "strands-agents>=1.18.0",
+    "strands-agents-evals>=0.1.0",
 ]
 
 [project.optional-dependencies]
 strands-agents = [
     "strands-agents>=1.1.0"
 ]
+strands-agents-evals = [
+    "strands-agents-evals>=0.1.0"
+]

src/bedrock_agentcore/evaluation/__init__.py

@@ -0,0 +1,19 @@
+"""AgentCore Evaluation integration for Strands."""
+
+from bedrock_agentcore.evaluation.integrations.strands_agents_evals.evaluator import (
+    StrandsEvalsAgentCoreEvaluator,
+    create_strands_evaluator,
+)
+from bedrock_agentcore.evaluation.span_to_adot_serializer import (
+    convert_strands_to_adot,
+)
+from bedrock_agentcore.evaluation.utils.cloudwatch_span_helper import (
+    fetch_spans_from_cloudwatch,
+)
+
+__all__ = [
+    "create_strands_evaluator",
+    "StrandsEvalsAgentCoreEvaluator",
+    "convert_strands_to_adot",
+    "fetch_spans_from_cloudwatch",
+]

src/bedrock_agentcore/evaluation/integrations/__init__.py

@@ -0,0 +1 @@
+"""AgentCore Evaluation integrations."""